[webgpu] Fused GeneratePositionIDs into FusedQKRotaryEmbedding #26335
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
xiaofeihan1
force-pushed
the
xiaofeihan/opt_generate_id
branch
from
0fb1cd4 to
ec3292b
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
This PR integrates GeneratePositionIDs into FusedQKRotaryEmbedding, improving overall performance.
Motivation and Context
Previously, for GQA, the processing flow was:
SplitPackedQKVProgram -> GeneratePositionIDs -> FusedQKRotaryEmbedding -> FlashAttentionAfter this change, the pipeline becomes:
SplitPackedQKVProgram -> FusedQKRotaryEmbedding -> FlashAttentionBy fusing GeneratePositionIDs into FusedQKRotaryEmbedding, we reduce kernel launches and memory operations.
On NV5080, generation TPS improves by ~4% (from 135.6 tps to 141.2 tps).